Table 4 Mediation analysis on the association between childhood maltreatment and T2D via BMI.

From: Childhood maltreatment, adulthood obesity and incident type 2 diabetes: a retrospective cohort study using UK Biobank

 

Model 1

Model 2

Estimate

LCIa

UCIb

p-value

Estimate

LCIa

UCIb

p-value

Total effect, HR

1.16

1.14

1.18

<0.001

1.17

1.14

1.20

<0.001

Total natural direct effect, HR

1.11

1.09

1.13

<0.001

1.12

1.09

1.14

<0.001

Pure natural indirect effect, HR

1.05

1.04

1.05

<0.001

1.05

1.04

1.05

<0.001

% of excess risk attributed to:

 Interaction with BMI

9.7

1.1

19.0

0.02

8.4

0.8

17.0

0.03

 Mediation via BMI

32.4

27.6

39.0

<0.001

31.1

26.0

36.1

<0.001

 Total via BMI

42.2

32.8

52.1

<0.001

39.4

32.9

49.2

<0.001

  1. Model 1 adjusted for age, sex, ethnicity, deprivation, education; Model 2 additionally adjusted for physical activity, TV viewing, sleep duration, alcohol consumption and smoking status as exposure-induced mediator-outcome confounders.
  2. aLCI - Lower 95% confidence interval.
  3. bUCI - Upper 95% confidence interval.
Back to article page